Enhancing Usability of Network-based Library Information System
- Experimental Studies of a User Interface for OPAC and of a Collaboration
Tool for Library Services
Shigeo Sugimoto, Seiki Gotou, Yanchun Zhao, Tetsuo Sakaguchi, Koichi Tabata
University of Library and Information Science (ULIS)
1-2, Kasuga, Tsukuba, Ibaraki 305, Japan
email: sugimoto, gotou, zhao, saka, firstname.lastname@example.org
KEYWORDS: User-friendliness of Library Information System,
Graphical User Interface, Collaboration Support System, Reference Work,
User-friendliness of library information systems is an important
consideration in the development of the digital library. This paper
describes two experimental studies of library information system usability:
Graphical User Interface (GUI) oriented OPAC (Online Public Access Catalog)
with a bookshelf-like browser named SOPAC, and a Collaboration Support
System (CSS) and its application to a reference task. SOPAC uses a bookshelf
metaphor where books are displayed and the image of each book is based on
its size. SOPAC accesses the OPAC database of the university library of ULIS
without any modification of that database. The CSS includes TV-phone, Shared
Virtual Display, Image Tool, and White Board. An experiment applying it to a
task teaching how to use a GUI-based OPAC via network is described. Both
systems have been developed on a conventional workstation LAN (Ethernet, 10
User-friendliness of library information systems is important, because they
are used not only by professional librarians but also by novice users.
Advanced hardware and software technologies have brought us new concepts to
build user-friendly library information systems such as GUIs, multimedia
user interfaces, collaboration support tools, etc. For example, electronic
document delivery systems such as the project mercury at CMU[mercury],
RightPages[rightpages], TULIP[tulip], and
NACSIS-ELS[adachi1][adachi2] provide users with GUIs to retrieve
and read documents. Large digital library projects such as those granted
by NSF will build a vast and useful information retrieval environment[nsf].
The digital library will provide us with an intellectual environment where
users can interactively retrieve and read documents, books, audio-visual
materials, etc. However, from a practical point of view, we cannot build
such environments in one day, because not all libraries can afford
high-performance computing and network facilities. In addition, since
library information systems are data-centered, it is not easy to modify
data-dependent functions or to add new information to an existing database
in order to create new functions and user interfaces. These points imply
that we need to build toward the digital library gradually, starting from
existing library information systems.
In this paper, we describe two experimental systems developed on a
conventional workstation LAN. The first system is an OPAC called SOPAC which
has a bookshelf-like browser. The second is a CSS designed for collaborative
library services between a library user and a librarian.
SOPAC has been developed on the database of an OPAC system called XOPAC
that is in use at the university library of ULIS[take]. A user can issue
retrieval commands and receive textual results through various windows of
SOPAC. He/she can show a bookshelf window where books are sorted in the
order of books on the physical bookshelves in the library. In the bookshelf
window, each book is represented as a book-shaped image which is a
push-button to request detailed information about the book. The book-shaped
image is created only from a record of the existing catalog database, i.e.,
title, author, and size. Users can see the height and thickness of a book in
addition to its title and author(s).
Collaboration is an important issue for the digital library, because this
library will be the principal environment for intellectual activities: a
user seeking information will consult with a librarian, and groups of users
will meet in a virtual conference room in the digital library using library
materials. The authors have developed a collaboration support system (CSS)
in order to evaluate the applicability of CSCW technologies to reference
services. The CSS includes a TV-phone, a Shared Virtual Display, an Image
Tool, and a White Board, all of which are considered essential for
communication between the librarian and the user. A CSS has been developed
on conventional UNIX workstations with audio-visual interfaces connected by
Ethernet (10 Mbps) LAN. We performed an experiment on the applicability of
the CSS to the task of teaching how to use XOPAC. Since XOPAC is GUI-based,
the instructor had to show the student dynamic features such as transition
of windows, movement of the mouse cursor and key input in a window. The
results with a sample of 30 students was positive.
ENHANCING USABILITY OF LIBRARY INFORMATION SYSTEMS
Information Access Support Tools in Libraries
Consider a user in a typical library. He/she would search for a book using
an OPAC terminal, find a book and go to a bookshelf designated by the OPAC.
Then, he/she would locate the book on the bookshelf and pick it up to browse
it. In some case, he/she would find another book which is more suitable to
his/her requirements than the retrieved one. In the case he/she cannot find
any book suitable to his/her requirements at the bookshelf, he/she would go
back to the OPAC terminal or consult with the reference librarian concerning
The preceding paragraph describes the typical behavior of a user in a
conventional library. The OPAC and the librarian are essential aids for the
user to access books in the library. From these considerations, the authors
found the following aspects to be important to improve the information
environment based on exisiting library information systems;
From these considerations, the following guidelines for OPAC usability
improvement were derived.
- Bridging the gap between OPAC and bookshelves: A user often goes back and
forth between an OPAC terminal and bookshelves, so that a tool like a
bookshelf browser coupled with an OPAC will be helpful.
- Improving accessibilty to a librarian: It is crucial especially for
remote users to have improved communication tools for collaborating with the
librarian. This is because, even if OPACs and electronically accessible
materials are well-provided, the users need specialized assistance.
It seems obvious that a collaboration environment for a library will be
implemented using a well-designed CSCW tool on a high-performance
workstation on a high-performance network. However, from a practical
viewpoint, we should examine the feasibility of collaborative reference work
in a conventional environment consisting of conventional workstations,
personal computers and network.
Existing databases should be used without modification.
Graphical (or multimedia) images should be used to help users browse
simulated bookshelves as well as physical bookshelves.
Bookshelves should be (dynamically) re-organizable in accordance with
Multiple OPACs including personalized ones should be accessible through a
single user interface (or terminal).
Active information, such as "this book is checked out", should be
iconized and given to users as well as bibliographic information.
User interface of OPAC should be adaptable to the user environment, e.g.
terminal type and user preferences.
Enhancing Usability of OPAC
OPAC has been the most common tool for library users and librarians, and it
will be also commonly used in digital libraries. It is obvious that
well-designed GUI improves user-friendliness especially for novice users.
OPAC with a bookshelf-like browser of bibliographical data would be useful
from a practical point of view, even if the bookshelf on a display does not
look like a physical bookshelf. The advantages of the bookshelf-like
interface are: 1) users can find not only a book retrieved via retrieval
commands but also from a glance at other books placed on the same bookshelf,
2) the structure of bookshelves is semantically organizable, and 3) the
display order of books is modifiable as needed.
HOPAC[hopac] has a bookshelf-like interface organization based on the Dewey
Classification System. Its hierarchical structure helps a user move his/her
viewpoint in the space of books and documents. The Book Shelf window of
HOPAC does not look like a physical bookshelf. In addition to the text-based
presentation, the authors consider that images of books and bookshelves
would help users find books. The number of book images displayable on a
display at a same time is obviously smaller than the number of book title
texts. This fact looks disadvantageous, but too much information on a
display is not friendly to users. The image-oriented bookshelf has the
advantage that users can immediately see the size of a book without reading
detailed bibliographic information. In addition, we can embed icons into a
book image which represent such information as the book being checked-out,
having duplications, one of series books, and so on.
Collaboration by Library Users and Librarian
The librarian is an important information resource in a library, and we
cannot fully take advantage of a library without his/her help. A digital
library will not necessarily be a fully automated librarian-less library,
but will have to provide intelligent collaborative services of librarians.
Yamamoto mentioned the roles of librarians in the era of digital
libraries as follows[yamamoto],
He also mentioned the types of reference services via network as follows,
Gateway between (novice) users and digital library networks,
Reference via networks,
Read-aloud via networks, and
Intermediary as an information creator, editor, and publisher.
Network Navigation: support to find and approach information sources,
Cooperative Search: support to search information,
Cooperative Translation: multi-linguistic support to access foreign information, and
Instruction of Information Tools: support to learn what tools users should
use and how they should use the tools.
Figure 1: Snapshot of SOPAC.
Figure 2: Bookshelf Window.
In addition to the collaborative services of librarians, the digital library
will have to provide environments for group work[lotus][marshall]. For example,
researchers sometimes use a conference room in a library for research
meetings using journals
and other materials in the library. They store their new knowledge
and information which are got from their discussion into the
library. This means that the digital library should provide a virtual
conference room where users can hold discussions through the same displays
where library materials such as electronic journals and digitized images are
Thus, collaboration support is one of the important features that should be
explored in order to build the digital library. Reference services could be
carried out using conventional media such as telephone, facsimile and email.
However, these media do not offer the face-to-face communication that is
quite important both for librarians and users; librarians must be able to
correctly understand questions and needs of users, and users would be able
to appropriately express their questions in their conversation with
OPAC WITH BOOKSHELF-LIKE BROWSER: SOPAC
SOPAC has been implemented based on the client-server model on a unix
workstation which has the Motif GUI environment. The database of SOPAC is
identical to the database used by XOPAC, which has 110,000 titles.
Users communicate with SOPAC via multiple windows. A snapshot of the SOPAC
GUI is shown in Fig 1. This GUI includes the following windows:
Retrieval Command Window -- accepts retrieval commands as well as
set operations AND, OR, and DIFF for intersection, union, and difference
operations on retrieved data.
Result Window -- displays retrieved data. Users can get detailed
bibliographic data of a member of the list on the Detailed Information
Memopad Window -- for user annotation.
Bookshelf Window -- browse a portion of bookshelf by designating a
book on a result window. A bookshelf image of the designated book and its
neighboring books is displayed. The order of the book images is primarily
determined by the identification numbers of the books.
Detailed Information Sub-window -- all of the bibliographic
information of a book which is selected on a result window or a bookshelf
Users can browse a bookshelf by clicking the Display Bookshelf button to
open the window shown in Fig.2. SOPAC displays a bookshelf with the book
highlighted on the Retrieval Window and placed in the center. The order of
books is the same as that on a physical bookshelf. The order is determined
by the unique identification number given to each book, and users can scroll
the bookshelf horizontally.
Each book image is a push-button to request detailed bibliographic
information, which is displayed on a Detailed Information Sub-window
attached below the bookshelf window. The shape of a book image is determined
by the bibliographic description of the book, i.e., height and number of
pages. The book title is presented from top to bottom and can be vertically
It is not easy to re-create an OPAC database or to modify the scheme of its
records. SOPAC has been developed on the XOPAC database. It would be better
for bookshelf-like browsers to be able to show physical images of books,
i.e., covers and title pages. However, it was not practical to scan in the
physical image of every holding and to add it to an existing OPAC database,
so the SOPAC book image is created only from existing data.
Books are displayed vertically in order to enhance realism. This means that
title texts on the images are drawn vertically. This is convenient for
titles in Japanese.
Several issues remain unsolved in the current SOPAC system:
Bookshelf reorganization: functions to re-organize bookshelves in
accordance with semantical requests from users,
Browsing Support Tools based on Semantical Structure of
Bookshelves: tools to show classification hierarchy and thesauri, and
Active Information Tools: icons to show active information such as
"checked out", "in bindery", etc.
COLLABORATION SUPPORT SYSTEM AND ITS APPLICATION TO A LIBRARY SERVICE
In this section, we firstly discuss a model of reference work, and then
describe the system structure of the CSS. Some experimental results using
the system are also described.
Modeling Reference Work as Librarian's Desk
The principal components of reference work in conventional libraries are as
Based on the objects listed above, we can model an environment for CSS. The
library user and librarian are modeled as the participants in a conversation
carried out via the system. Library materials and user materials, which may
or may not be in digital form, are modeled as objects to be displayed on the
terminals in front of the participants. The librarian's desk is modeled as
a display screen in front of each participant where the participants can see
their partners and view the same materials.
a library user,
library materials, which include primary information sources in a
library such as books, journals, etc., and secondary information sources
such as OPACs, thesauri, dictionaries, and encyclopedias,
user's materials such as books and memos, and
a librarian's desk.
It is natural to design a library collaboration support system based on the
librarian's desk metaphor. We can find memo pads, a computer terminal, a
telephone, a copy machine, and so on in addition to the library materials
on/around a librarian's desk. These things should be modeled as system
components visible to the participants. For example, the participants would
enhance their discussion with visual materials and memos displayable on
Figure 3: System Overview.
Figure 4: Snapshot of the Collaboration Support Tool(TV-phone and SVD)
Figure 5: Lan Environment.
- The librarian (L) explains a retrieval window.
- L shows a simple retrieval operation to the user (U).
- L shows and explains the result window
where the retrieved result is displayed.
- L copies the contents of the result window into a Memopad window,
and explains its usage.
- L explains and shows an example of a composite retrieval which implies
AND, OR, and NOT operation.
- L shows the result.
- L explains the operation to terminate the OPAC.
- U does a simple retrieval
and displays its result with assistance by L.
- U uses a scratch window with assistance by L.
- U does a composite retrieval and displays
its result with assistance by L.
- L asks U whether U has any questions about the usage.
- L tells U to do practice until U gets satisfied.
- U says Good-bye.
(Note: The OPAC has a retrieval window to input retrieval commands,
result wind ows to display retrieval results,
and scratch windows to make personal annotati ons.)
Figure 6: Retrieval Instruction Scenario
- SONY NEWS-3000 workstations
CPU: MIPS R3000, 20/25 MHz, 32/64 MB RAM, 17/25 MIPS
Display: 1024(W) x 768(H), 15 bits true color (32767 colors)
Video Input: NTSC
LAN interface: Ethernet
Window System: X-window Version 11, Release 6 OSF/Motif
- Backbone LAN: FDDI (100 Mbps)
Front-end LAN: Ethernet (10 Mbps).
- Audio Sound
- Sampling Rate: 8 KHz
Quantumization: 8 bits
Unit Data Size: 8 KByte (0.4 sec/unit)
- Video Image
- Capturing Rate: 2 frames/second
Size: 160(W) x 120(H)
Image Size: 38 KByte/frame
Table 1. Hardware Specifications used in Experiment
The CSS has been developed on two workstations of the ULIS campus LAN. The
graphical environment of the workstation is OSF/Motif on X-window (X11R6).
The overview of the system on each workstation is shown in Fig.3. Each
component is realized as a window. A snapshot of the terminal display is
shown in Fig.4. The following paragraphs are brief descriptions of the
Overview of the LAN environment and specifications of hardware used in this
study are shown in Fig. 5 and Table 1, respectively.
Control Tool: The control tool manages initiation and termination of
a session between a user and a librarian.
TV-phone: The TV-phone displays real time video pictures taken by a video camera
placed at and digitized by the workstation. The captured image is digitized
and transferred through the LAN. Sound communication is half-duplex,
requiring a talk control panel to switch talking direction.
Shared Virtual Display (SVD): SVD is a window which realizes a virtual display
where multiple windows are displayed and input/output operations are performed
on them. The participants share the virtual display where the user can see every
operation by the librarian, and vice versa. This feature is crucial in
order to enhance the quality of communication between the participants; that
is, a participant needs to view all operations performed on the terminal of
the other participant, for example mouse movement, mouse clicking, etc.
The participants can input character text and draw images on the board which
is equally accessible to both participants. The participants can use the
white board to explain things which are difficult to express on the TV-phone.
The participants can input images using a color image scanner (max. 400
dpi). The input images are shown on both workstations.
Applying Collaboration Support Tool to Library Service
We applied the CSS to the task written below.
TASK: A user who wants to use a GUI-based OPAC requests and receives instruction
from a librarian. Librarian teaches him/her how to use the OPAC.
The user is assumed to have basic knowledge about OPAC and to have some
computer experience, not necessarily with GUI. The librarian (instructor)
shows how to use the mouse and keyboard to input parameters and invoke
operations. The user (student) learns and practices the example operations
given by the instructor. The windows on SVD accept equally keyboard and
mouse inputs from both participants. The mouse cursors of each participant
are visible to both participants.
The scenario given to the librarian for the experiment is shown in Fig. 6.
Since this scenario involves interactions which are common among interactive
services by librarians, we can evaluate the applicability of the system not
only to the service written above but also to interactive services in general.
Experiment and Evaluation
30 users performed the task in the scenario given above using two remotely
placed workstations. They were asked to fill out a questionaire shown below
before and after the experiment. For comparison, four users received
instruction in the same task from an instructor sitting next to the user
in front of a workstation.
Questionaire (Definitely Yes: 5 - Definitely No: 1)
All of the participants were students of ULIS. About 20 % were novice users
of workstations, not used to a mouse of GUI-based applications. All
students had learned information retrieval with the ULIS-OPAC on a
main-frame with a character user interface. The OPAC used in this
experiment, called XOPAC, is a new one, so that not all of the students had
- [Q.1]Are you used to use keyboard and mouse ?
- [Q.2]Was the talk button easy to use ?
- [Q.3]Was it easy for you to hear the librarian ?
- [Q.4]Did you smoothly communicate with the librarian via TV-phone ?
- [Q.5]Was it easy to understand the instructions pointed by the mouse cursor ?
- [Q.6]Was it easy to follow the librarians operations ?
- [Q.7]Did the librarian understand what you told through TV-phone ?
- [Q.8]Do you like to learn OPAC via this collaborative support system better than to learn OPAC just beside the librarian using one terminal ?
- [Q.9]Did TV-phone and SVD work together well ?
- [Q.10]Do you think this kind of system is advantageous under LAN environment ?
question 5 4 3 2 1 total
Q.1 6 10 8 2 4 30
Q.2 7 10 8 4 1 30
Q.3 1 5 8 14 2 30
Q.4 5 8 7 10 0 30
Q.5 16 12 1 1 0 30
Q.6 15 11 2 2 0 30
Q.7 12 10 6 2 0 30
Q.8 7 7 13 3 0 30
Q.9 8 16 5 1 0 30
Q.10 16 10 4 0 0 30
Table 2: Responses to the Questionnaire
minutes | 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
persons | 1 0 2 0 5 2 4 4 4 2 3 2 0 0 0 1
Table 3: Minutes for Instruction by Librarian Using the Collaboration Support
Responses to the questionaire on usability of the system is summarized in
Table 2. Response to the system was mostly positive, and criticism was due
mainly to the low quality of the TV-phone, such as poor sound, image size
and frame rate, and half-duplex communication. Since such criticism was
mainly of the performance rather than structure, these problems can be
solved in future high performance network and computers.
minutes | 10 11
persons | 2 2
Table 4: Minutes for In-person Instruction
Instruction time required using the network CSS is summarized in Table 3,
and using in-person instruction in Table 4. The average instruction times of
the CSS and the in-person groups are 16.8 and 10.5 minutes, respectively.
The fastest and slowest times of the CSS group are 10 and 25 minutes,
respectively. The instruction times of 26 CSS users (87%) are between 14 to
21 minutes, and the median is 17 minutes. The instruction time of the CSS
group has a wide range compared with the four students of the in-person
We attribute the difference betweenthe two groups to the fact that the CSS
students had to learn two systems, the OPAC and the CSS, since the
experiment was the first experience to use the CSS for all of them, and some
were novice users of workstations.
We feel that the difference in instruction times between the two groups is
reasonable and acceptable, because all users learned to use the OPAC without
going to the librarian's desk at a cost of only 7 additional minutes.
`Digital library' is a fascinating term, but many problems remain to be
solved and a high-performance computing and communication environment is
needed to achieve the ideal digital library. Nevertheless, in this paper, we
described two systems developed in a conventional environment which indicate
that the usability of library information systems will be greatly enhanced
by the future digital library.
[adachi1]Adachi, J., Premises for Establishment of an Electronic
Library System, Digital Libraries (ISSN 1340-7287), No.1, 1994.8 (in Japanese)
(http://www.dl.ulis.ac.jp/DLW_E/Journals/no1.html, Abstract in English)
[hopac] Allen, R.B., Navigating and Searching in Hierarchical Digital
Library Catalog, Proc of DL'94, pp. 95 - 100, College Station, TX, 1994.6
[lotus] Ehrlich, K. and Cash, D., Turning Information into Knowledge:
Information Finding as a Collaborative Activity, Proc. of DL'94, pp. 119 -
125, College Station, TX, 1994.6
[take] Fujita, T. et al., Transporting an Online Public Access Catalog
from Mainframe to Distributed Environment, Proc. of 47th FID, pp.444-448,
[rightpages] Hoffman, M.M., et al., The RightPages Services: An
Image-Based Electronic Library, Journal of the American Society for
Information Science, Vol.44, No.8, pp.446-452, 1993.9
[marshall] Marshall, C.C. et al., Putting Digital Libraries to Work:
Issues from Experience with Community Memory, Proc. of DL'94, pp. 126 - 133,
College Station, TX, 1994.6
[mercury] Trol, D.A., Project Mercury and Development of the Library
Information System, Mercury Technical Report Series, No.7, 1993
[yamamoto] Yamamoto, T., The Digital Librarian: Work and Tools, Digital
Libraries (ISSN 1340-7287), No.1, Tsukuba, 1994.8 (in Japanese)
(http://www.dl.ulis.ac.jp/DLW_E/Journals/no1.html, Abstract in English)
[tulip] ELSEVIER SCIENCE B.V., TULIP - The University Licensing Program
http://www.elsevier.nl/info/projects/tulip.htm, March 1995.
[nsf] National Science Foundation, NSF ANNOUNCES AWARDS FOR DIGITAL
LIBRARIES RESEARCH, http://www.nsf.gov/nsf/press/pr9452.htm, NSF-PR 94-52, Sept. 1994
[adachi2] The National Center for Science Information Systems (NACSIS),
NACSIS Digital Library Project, http://www.nacsis.ac.jp/dl-j.html (in Japanese)